Polarization

class pyvisgen.simulation.Polarization(SI: tensor, sensitivity_cut: float, amp_ratio: float, delta: float, polarization: str, field_kwargs: dict, random_state: int, device: device)

Bases: object

Simulation of polarization.

Creates the \(2\times 2\) stokes matrix and simulates polarization if polarization is either 'linear' or 'circular'. Also computes the degree of polarization.

Parameters:

SItensor()

Stokes I component, i.e. intensity distribution of the sky.

sensitivity_cutfloat

Sensitivity cut, where only pixels above the value are kept.

amp_ratiofloat

Sets the ratio of \(A_{X|R}\). The ratio of \(A_{Y|L}\) is calculated as 1 - amp_ratio. If set to None, a random value is drawn from a uniform distribution. See also: random_state.

deltafloat

Sets the phase difference of the amplitudes \(A_{X|R}\) and \(A_{Y|L}`\) of the sky distribution. Defines the measure of ellipticity.

polarizationstr

Choose between 'linear' or 'circular' or None to simulate different types of polarizations or disable the simulation of polarization entirely.

random_stateint

Random state used when drawing amp_ratio and during the generation of the random polarization field.

devicedevice

Torch device to select for computation.

Methods Summary

circular()	Computes the stokes parameters I, Q, U, and V for circular polarization.
dop()	Computes the degree of polarization for each pixel.
linear()	Computes the stokes parameters I, Q, U, and V for linear polarization.
rand_polarization_field(shape[, order, ...])	Generates a random noise mask for polarization.
stokes_matrix()	Computes and returns the 2 x 2 stokes matrix B.

Methods Documentation

circular() → None

Computes the stokes parameters I, Q, U, and V for circular polarization.

This is done using the following equations:

\[\begin{split}I &= A_R^2 + A_L^2 \\ Q &= 2A_R A_L \cos\delta_{RL} \\ U &= -2A_R A_L \sin\delta_{RL} \\ V &= A_R^2 - A_L^2\end{split}\]

dop() → None

Computes the degree of polarization for each pixel.

linear() → None

Computes the stokes parameters I, Q, U, and V for linear polarization.

This is done using the following equations:

\[\begin{split}I &= A_X^2 + A_Y^2 \\ Q &= A_X^2 - A_Y^2 \\ U &= 2A_X A_Y \cos\delta_{XY} \\ V &= -2A_X A_Y \sin\delta_{XY}\end{split}\]

rand_polarization_field(shape: list[int, int] | int, order: list[int, int] | int = 1, random_state: int = None, scale: list = None, threshold: float = None) → tensor

Generates a random noise mask for polarization.

Parameters:

shapearray_like (M, N), or int

The size of the sky image.

orderarray_like (M, N) or int, optional

Morphology of the random noise. Higher values create more and smaller fluctuations. Default: 1.

random_stateint, optional

Random state for the random number generator. If None, a random entropy is pulled from the OS. Default: None.

scalearray_like, optional

Scaling of the distribution of the image. Default: [0, 1]

thresholdfloat, optional

If not None, an upper threshold is applied to the image. Default: None

Returns:

imtorch.tensor

An array containing random noise values between scale[0] and scale[1].

stokes_matrix() → tuple

Computes and returns the 2 x 2 stokes matrix B.

Returns:

Btorch.tensor

2 x 2 stokes brightness matrix. Either for linear, circular or no polarization.

masktorch.tensor

Mask of the sensitivity cut (Keep all px > sensitivity_cut).

lin_doptorch.tensor

Degree of linear polarization of every pixel in the sky.

circ_doptorch.tensor

Degree of circular polarization of every pixel in the sky.